Neuropathologic studies show inflammatory molecules and activated microglia in the brain in AD. Epidemiological studies suggest that NSAID use lowers the risk of Alzheimer's Disease (AD). Risk reduction by NSAIDs could be due to anti-inflammatory actions, mediated by inhibition of cyclo-oxygenase 2 (COX 2). However, NSAIDs have many biological actions. This project stems from a novel action: many NSAIDs lower levels of Abeta42 in medium from cultured cells and in the brains of transgenic mice that over-express mutant APP. Selective Co2-2 inhibitors do not appear to lower Abeta42, suggesting that this effect is Cox-2 independent. Abeta42 is a key molecule in AD. It is overproduced in early onset genetic forms of AD, is deposited extensively in the brain in AD, and may initiate a cascade that results in neurodegeneration. Project by Galasko will use biological markers to characterize these Abeta-lowering effects in humans. In a series of studies, selected NSAIDs will be given to groups of patients with AD for two week periods. Levels f Abeta42 will be measured in CSF and plasma, and compared to a group of patients who receive a placebo. Parallel studies will also be conducted in non-demented elderly individuals. Biological markers related to neurodegeneration 9tauy and phospho-tau) oxidative damage (isoprostanes) and inflammation (chemokines) will also be measured in CSF. These biomarker studies will examine the Abeta- lowering effect in humans, and its relationship to different types of NSAIDs, low and high doses, and to other mechanisms of NSAID action. NSAID effects will then be investigated over 48 weeks. In a double blind, placebo-controlled study, 60 patients with mild AD will receive either the most promising NSAID or an identical placebo. Clinical outcome measures (cognitive and functional abilities) and biomarkers in plasma and CSF will be measured.